# Credit Attribution Codebook

## 1. FIRST CHECK: Is there any credit claim? (gives_credit)

Decision tree for gives_credit:
1. Does the statement contain explicit credit through:
   a) Causal verbs (enabled, secured, funded), OR
   b) Attribution of decision-making ("contributes to our decision", "influenced by"), OR
   c) Economic environment claims ("thanks to", "because of", "due to")
   → If YES to any, set gives_credit=1 and continue to STEP 2
   → If NO, continue to question 2

2. Does the statement show implicit credit? Check ALL:
   - Actor attends/hosts ceremony for specific project
   - Actor makes announcement about specific project AND frames it as achievement/success
   - Actor publicly associates with project success through active language
   → If YES to any, set gives_credit=1 and continue to STEP 2
   → If NO to all, continue to question 3

3. Is this merely descriptive/informative? Check ALL:
   - Technical specifications or equipment lists
   - Routine business updates or status reports
   - Factual job numbers without attribution
   - Standard press release boilerplate
   → If YES to any, set ALL variables to 0 and STOP
   → If unsure, look for active credit claiming language

Key Distinction: Credit claiming includes both direct causation ("funded by") AND influential attribution ("contributed to our decision"). When an actor cites something as influencing their decision or creating favorable conditions, this counts as giving credit.

Examples of NO credit:
- "New factory will create 500 jobs" (just reporting)
- "Company X announced plans to expand" (passive description)
- "Project includes following equipment..." (technical details)
- "The IRA sets ambitious goals" (just mentioning)

Examples of YES credit:
- "Our state attracted this investment" (active role)
- "Thanks to our business climate..." (explicit attribution)
- "We're proud to announce this success" (association with achievement)
- "The IRA's impact on industrial growth contributed to our decision" (economic environment attribution)
- "Growth connected to the IRA influenced our expansion plans" (policy impact attribution)

## 2. WHO gets credit? (Only check if gives_credit=1)

### Social Media Attribution Rules
Social media posts require special attention for credit attribution:

1. @mentions and Tags:
   - Credit when mentioned in success/achievement context
   - Credit when listed as partner/collaborator
   - NO credit for mere cc's or FYI mentions
   - NO credit for requests or complaints

2. Partnership Language:
   - "partnership of @Actor" = credit to Actor
   - "working with @Actor" = credit to Actor
   - "thanks to @Actor" = credit to Actor
   - Must be in context of success/achievement

3. Verification Requirements:
   - Verify correct role (U.S. Senator vs State, etc.)
   - Check handle matches official account

Examples:
YES credit:
- "Partnership of @POTUS in this success" → credit_biden=1
- "Working with @SenatorName to deliver" → credit_senate=1
- "Thanks to @RepName's leadership" → credit_us_rep=1
- "@GovName helping create jobs" → credit_governor=1

NO credit:
- "cc: @Actor" (just FYI)
- "@Actor please help" (request)
- "Following up with @Actor" (process)
- "@Actor what's the status?" (question)

### Federal Actors

credit_biden=1 if:
- Names President/White House OR official handle (@POTUS, @WhiteHouse)
- Shows credit through:
  * Explicit verbs (secured, helped)
  * Ceremonies/events
  * Partnership language
  * Joint announcements
- NO credit for general "federal" references

credit_senate=1 if:
- Names specific U.S. Senator (NOT state senate) OR official handle
- Shows credit through:
  * Explicit verbs (secured, helped)
  * Ceremonies/events
  * Partnership language
  * Joint announcements
- First-person credit if speaker is U.S. Senator

credit_us_rep=1 if:
- Names U.S. Representative OR official handle
- Shows credit through:
  * Explicit verbs (secured, helped)
  * Project announcements/promotion
  * Ceremonies/events
  * Partnership language
- NO credit for just being cc'd/contact

### State & Local Actors

credit_governor=1 if:
- Names Governor OR official handle
- Shows credit through:
  * Explicit verbs
  * Ceremonies/events
  * Partnership language
  * Joint announcements
- "My administration" if Governor speaking

credit_local=1 if:
DOES count:
- Financial support/incentives
- Active recruitment
- Explicit thanks/credit
- Ceremonies/announcements
- Partnership language

Does NOT count:
- Regular permits/zoning
- Land transactions
- Administrative processes

### Party & Law Credit

credit_dem/credit_gop=1 if:
- EXACT party name appears ("Democrat(s)", "Republican(s)")
- Shows causal connection
- NO credit for just (D) or (R) labels

credit_ira/credit_bil=1 if:
- Direct funding/incentives used
- Explicitly aided decision
- Broader economic impact cited as factor in decision
- Industry growth effects attributed to law
- NO credit for just goals/eligibility

## 3. Credit Attribution Language Guide

### Direct Causal Language
Words that explicitly show enabling or causing:
- enable, secure, fund, finance
- make possible, help land
- deliver, create, bring
- authorize, approve
- invest in, support

### Decision Influence Language
Words that show impact on choices:
- contributes to (our decision)
- influenced by
- due to, because of
- thanks to, credit to
- in response to

### Partnership Language
Words that show active involvement:
- partnership with, working with
- collaboration between
- joint effort with
- together with
- alongside

### Ceremony/Event Language
Words that show public association:
- announce, unveil
- celebrate, welcome
- break ground, cut ribbon
- host, attend (with project context)
- showcase, highlight

## 4. Calibration Examples

1. "Thanks to President Biden's leadership, we secured two billion dollars for this solar factory."
   → gives_credit=1, credit_biden=1
    [Explicit credit with causal verb]

2. "This project meets IRA ten percent bonus criteria."
   → all zeros
   [Mere eligibility without credit claim]

3. Governor quote: "We secured LG Chem's three billion dollar investment for our state."
   → gives_credit=1, credit_governor=1
   [Self-credit through first person + success claim]

4. "Senator Hickenlooper joined President Biden to cut the ribbon on the wind tower plant."
   → gives_credit=1, credit_biden=1
   [Implicit credit through ceremony]

5. Mayor: "Our city council worked for years to land this plant."
   → gives_credit=1, credit_local=1
   [Explicit local credit with causal connection]

6. "Federal leaders celebrated the opening."
   → all zeros
   [Generic reference without specific actor]

7. Company CEO: "Thanks to Governor Lee's business climate, we chose Tennessee."
   → gives_credit=1, credit_governor=1
   [Third-party giving explicit credit]

8. "Thanks to the Florence County Council for making this investment possible."
   → gives_credit=1, credit_local=1
   [Explicit local credit from company]

9. "The Bipartisan Infrastructure Law and the Inflation Reduction Act make transformational investments in transportation electrification. ABB E-mobility is expanding its US operations."
   → gives_credit=1, credit_ira=1, credit_bil=1
   [Broader economic impact cited as factor]

10. "The North Carolina Senate approved funding for the project."
    → all zeros
    [State senate reference - do not code as U.S. Senate]

## 5. Metadata Usage

Metadata lines that may appear before the ### delimiter:
- speaker: speaker or company name
- role: Company | Governor | U.S. Senator | U.S. Representative
- state: two-letter postal code (e.g., GA)
- district: House district (e.g., GA-07) if known
- city: city or county if helpful context
- release_type: primary | proxy | social
- ira_funding: YES | NO (YES means text says IRA funding or programs benefitted the project)
- bil_funding: YES | NO (YES means text says BIL funding or programs benefitted the project)

Key metadata rules:
- Self-credit applies when role matches actor (Governor, Senator, Representative) and first-person used
- In proxy releases, only quoted text can claim credit
- If ira_funding=NO, need explicit new funding language for credit_ira=1
- If bil_funding=NO, need explicit new funding language for credit_bil=1

## 6. Output Format

{
 "gives_credit":0,
 "credit_biden":0,
 "credit_senate":0,
 "credit_us_rep":0,
 "credit_governor":0,
 "credit_local":0,
 "credit_dem":0,
 "credit_gop":0,
 "credit_ira":0,
 "credit_bil":0
}

Every field must be present and set to 0 or 1.